gila_all <- read_csv("data_gmocc-all-nv.csv")
gila_tracks <- filter(gila_all, det_type_strict != "v")
gila_vis <- filter(gila_all, det_type_strict != "s")
ggplot(gila_all, aes(rock_ind, det_strict)) +
geom_point(size=3) +
geom_smooth(method="lm",
fullrange = TRUE) +
labs(title="Gila monster detections - all data") +
ylab ("Probability of Detection") +
xlab ("'Sandiness Index'")
## `geom_smooth()` using formula = 'y ~ x'
ggplot(gila_all, aes(rock_ind, det_strict)) +
geom_point(size=3) +
geom_smooth(method="glm",
fullrange = TRUE,
method.args=list(family="binomial"(link="logit"))) +
labs(title="Gila monster detections - all data") +
ylab ("Probability of Detection") +
xlab ("'Sandiness Index'")
## `geom_smooth()` using formula = 'y ~ x'
ggplot(gila_all, aes(rock_ind, det_strict)) +
geom_point(size=2, position=position_jitter(width=0.1, height=0.15)) +
geom_smooth(method="glm",
fullrange = TRUE, #force fit line past limits of data
method.args=list(family="binomial"(link="logit"))) +
labs(title="Gila monster detections - all data") +
ylab ("Probability of Detection") +
xlab ("Sandiness Index")
## `geom_smooth()` using formula = 'y ~ x'
model_gila_all <- glm(det_strict ~ rock_ind, data=gila_all, family=binomial)
model_gila_all
##
## Call: glm(formula = det_strict ~ rock_ind, family = binomial, data = gila_all)
##
## Coefficients:
## (Intercept) rock_ind
## -6.7959 0.7527
##
## Degrees of Freedom: 1476 Total (i.e. Null); 1475 Residual
## Null Deviance: 381.8
## Residual Deviance: 334.9 AIC: 338.9
newdata <- data.frame(rock_ind = seq(0, 7, length.out = 100))
newdata$logit_fit <- predict(model_gila_all, newdata = newdata, type = "link")
gila_all$det_adj <- (gila_all$det_strict + 0.5) / 2
gila_all$logit_obs <- log(gila_all$det_adj / (1 - gila_all$det_adj))
ggplot() +
geom_point(data = gila_all, aes(x = rock_ind, y = logit_obs),
color = "black", size = 2, alpha = 0.8) +
geom_line(data = newdata, aes(x = rock_ind, y = logit_fit),
color = "blue", linewidth = 1.2) +
labs(
title = "Gila monster detections - all data - Logit Scale") +
ylab ("Probability of Detection") +
xlab ("Sandiness Index")
x <- predict(model_gila_all)
y <- resid(model_gila_all)
binnedplot(x, y)
coef(model_gila_all)
## (Intercept) rock_ind
## -6.7958927 0.7526968
confint(model_gila_all)
## Waiting for profiling to be done...
## 2.5 % 97.5 %
## (Intercept) -8.2138570 -5.593209
## rock_ind 0.5145914 1.017941
summary(model_gila_all)
##
## Call:
## glm(formula = det_strict ~ rock_ind, family = binomial, data = gila_all)
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -6.7959 0.6662 -10.201 < 2e-16 ***
## rock_ind 0.7527 0.1280 5.882 4.05e-09 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 381.84 on 1476 degrees of freedom
## Residual deviance: 334.91 on 1475 degrees of freedom
## AIC: 338.91
##
## Number of Fisher Scoring iterations: 7
1.384/4=0.346*100= 34.6%
ggplot(gila_tracks, aes(rock_ind, det_strict)) +
geom_point(size=2, position=position_jitter(width=0.1, height=0.15)) +
geom_smooth(method="glm",
fullrange = TRUE, #force fit line past limits of data
method.args=list(family="binomial"(link="logit"))) +
labs(title="Gila monster detections - visual and tracks") +
ylab ("Probability of Detection") +
xlab ("Sandiness Index")
## `geom_smooth()` using formula = 'y ~ x'
model_gila_all <- glm(det_strict ~ rock_ind, data=gila_tracks, family=binomial)
model_gila_all
##
## Call: glm(formula = det_strict ~ rock_ind, family = binomial, data = gila_tracks)
##
## Coefficients:
## (Intercept) rock_ind
## -10.573 1.384
##
## Degrees of Freedom: 1464 Total (i.e. Null); 1463 Residual
## Null Deviance: 292.7
## Residual Deviance: 219.1 AIC: 223.1
x <- predict(model_gila_all)
y <- resid(model_gila_all)
binnedplot(x, y)
coef(model_gila_all)
## (Intercept) rock_ind
## -10.57333 1.38420
confint(model_gila_all)
## Waiting for profiling to be done...
## 2.5 % 97.5 %
## (Intercept) -13.2989936 -8.328649
## rock_ind 0.9853142 1.848857
summary(model_gila_all)
##
## Call:
## glm(formula = det_strict ~ rock_ind, family = binomial, data = gila_tracks)
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -10.5733 1.2652 -8.357 <2e-16 ***
## rock_ind 1.3842 0.2197 6.299 3e-10 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 292.69 on 1464 degrees of freedom
## Residual deviance: 219.11 on 1463 degrees of freedom
## AIC: 223.11
##
## Number of Fisher Scoring iterations: 8